Subject: RE: mdvawv try advancing again too<br/>
Date: Aug 19, 2004 @ 05:48<br/>
Author: Ron McConnell (&quot;Ron McConnell&quot; &lt;rcmcc@...&gt;)<br/>
<a href="15101.htm">Prev</a>&nbsp;&nbsp;&nbsp;&nbsp;Post in Topic&nbsp;&nbsp;&nbsp;&nbsp;<a href="15107.htm">Next</a>&emsp;[<a href="topic15088.htm">All</a> Posts]<br/>
<a href="15102.htm">Prev</a>&nbsp;&nbsp;&nbsp;&nbsp;Post in Time&nbsp;&nbsp;&nbsp;&nbsp;<a href="15104.htm">Next</a><br/>
<br/><br/>
<div id="ygrps-yiv-1462657580">UPDATE:<br/>
<br/>
(1) N 39º 19&#39; 16.80205&quot;  W 77º 43&#39; 10.14059&quot;<br/>
<br/>
(2) N 39º 19&#39; 16.17539&quot;  W 77º 43&#39; 10.75219&quot;<br/>
<br/>
I said,<br/>
&quot;I now notice a slight difference in the gcb output<br/>
from GCGC, 79 ft _9_in. [not 79 ft 6.76 in]<br/>
...lat/long seconds being rounded off<br/>
after 3 decimal places on input.&quot;<br/>
<br/>
Actually, seconds were being truncated, not rounded<br/>
after 2 decimal digits.  I quietly,<br/>
without changing the prompts &quot;N/S dd mm ss.ss&quot;<br/>
and &quot;W/E ddd mm ss.ss&quot; allowed up to 5 decimal<br/>
digits for seconds for gcb.  I haven&#39;t changed<br/>
the degrees and decimal minutes or decimal degrees input.<br/>
Now gcb and GCGC distance agree for Vincenty.<br/>
The various algorithm results are:<br/>
<br/>
Sphere G. C. 24.253 m  = 964.8425 in = 79 ft 6.8425 in<br/>
Vincenty     24.251 m  = 954.7638 in = 79 ft 6.7636 in<br/>
Sodano       24.251 m<br/>
&quot;US Navy&quot;    24.251 m<br/>
Andoyer      24.236 m  = 954.1732 in = 79 ft 6.1732 in<br/>
Rhumb Line   24.253 m<br/>
&quot;aletheiak&quot;  24.2511 m = 954.7695 in = 79 ft 6.7695 in<br/>
<br/>
&quot;Andoyer&quot; is an ellipsoidal algorithm touted<br/>
for accuracy, worldwide &quot;error on order of 50m.&quot;<br/>
But in my experience, it is less accurate<br/>
than a good spherical great circle solution as seen above.<br/>
I will go back and quadruple check my code someday,<br/>
but my results agree with the Meeus book example<br/>
to nearly 5 places, and agree to 6 places<br/>
with a newer online Andoyer calculator.<br/>
Another Andoyer offline calculator<br/>
that brags about accuracy,<br/>
only agrees to 4 places with the book.<br/>
It is even farther off from Vincenty, Sodano, et al.<br/>
So... don&#39;t believe &quot;them&quot; without some proof.<br/>
<br/>
BTW, I need to look up the number,<br/>
but WGS-84 and NAD-83 are equivalent<br/>
within a fraction of a millimeter.<br/>
NAD-27 (a.k.a. Clarke 1866) can be different<br/>
by a 300 yards (or 300 meters) for the same<br/>
latitude & longitude values as NAD-83.<br/>
I&#39;ve seen references to differences<br/>
of a half mile for _somewhere_ (?) in the US.<br/>
Most older North American maps use NAD-27.<br/>
Most property boundaries are defined<br/>
by NAD-27 if lat/long are given.<br/>
[Wait until the lawyers hear about that one.  :) ]<br/>
<br/>
The new gcb has been uploaded to my web site.<br/>
I am _not_ going to advertise <br/>
the &quot;seconds to 5 decimals&quot;<br/>
outside the BP group.  Allowing 2 decimal points<br/>
is stretching normal civilian capability quite a bit.<br/>
Usually I would quietly laugh if given<br/>
more than 1 decimal digit for seconds,<br/>
But... when the data source is the surveyor... OK.  <br/>
They had better be right.  <br/>
That&#39;s why they get the big US$.  :)<br/>
<br/>
Have fun.<br/>
<br/>
Cheers, 73,<br/>
<br/>
Ron McC.<br/>
<a rel="nofollow" target="_blank" href="mailto:w2iol@...">w2iol@...</a><br/>
<br/>
Ronald C. McConnell, PhD<br/>
<br/>
 WGS-84: N 40º 46&#39; 57.6&quot;  +/-0.1&quot;<br/>
         W 74º 41&#39; 22.1&quot;  +/-0.1&quot;<br/>
         FN20ps.77GU31 +/-<br/>
 NAD-27  V +5058.3438  H +1504.2531<br/>
<br/>
 <a rel="nofollow" target="_blank" href="http://home.earthlink.net/~rcmcc">http://home.earthlink.net/~rcmcc</a> <br/>
<br/>
<br/>
pi approx. = 355/113<br/>
   (7 digit accuracy with 6 digits)</div>